The Agent Skills Directory

[SAFE]: No malicious patterns were detected. The skill focuses on structured software testing, emphasizing test independence and automated environment cleanup.
[DYNAMIC_EXECUTION]: The skill includes an explicit 'Import path policy' that prohibits the use of sys.path.insert(...) in individual test modules. This is a security-positive instruction that avoids ad-hoc path manipulation, favoring centralized and predictable import management through conftest.py files.
[INDIRECT_PROMPT_INJECTION]: The skill enables the creation of tests that ingest data from potentially untrusted external sources such as MQTT brokers, REST APIs, and web interfaces.
Ingestion points: tests/functional/ and tests/ui/ files processing network payloads and UI elements.
Boundary markers: Absent in the provided authoring patterns, although logical isolation is provided by the scenescape_env fixture.
Capability inventory: Operations include network requests via RESTClient, MQTT interactions, and containerized stack management via the docker backend.
Sanitization: Not explicitly mentioned for ingested test data, which is typical for functional testing. The surface area is low-risk and consistent with the primary purpose of exercising a networked software stack.

testing